Foreign Languages Lose Out in the Facebook vs. Google War
Today a friend of mine wrote a happy new year message in Chinese on her Facebook. I noticed a fairly new feature – Bing in-context translation. I clicked it and got this:
“United States Central time new year’s day. We wish you all the best of the new year, research on sophisticated, a happy horse!”
Not a very good translation, as you can see.
As a lover of learning foreign languages, and the mistranslations that can ensue, I’ve been playing around on translation software since the beginning of this century. For the most part, it was laughably erroneous, with AltaVista’s Babelfish tool giving me countless pages of hilarious nonsense.
Since Google translate has come around, I’ve started to believe that software can actually help you learn a language the way it’s used, not the way a bunch of algorithms lay out. Language is itself not always mathematical, or perfect.
Google translate is interesting in that its creator, Franz Josef Och, doesn’t believe in the effectiveness of rule-based algorithms. Instead, they use a comparative system drawn from a large database of texts translated into every language, which they collected from the UN. This system translates the texts into English and back, meaning you eliminate the necessity for collections like “Swahili into Norwegian.” This system isn’t perfect, mostly since English is a wacky language, but they have more human controls that have allowed for suggested translations, as well as drop-down alternative translations. The overall effect is a lot more human, and thus more accurate.
Bing’s translator is nothing to write home about (or to write in foreign languages with).
I know Google and Facebook are rivals, but it’s disappointing that Facebook can’t utilize Google’s translator to help friends speak between languages. Facebook as a place to translate social messages in context could be a huge opportunity for language learning, and it’s too bad Microsoft’s underwhelming software had to be the go-to choice.